The rapid pace at which GenAI tools are changing the world doesn’t leave us time to reflect on or even process it.
To ruminate on those complex societal issues, here we reflect and shelve.
Hopefully, this will give our reflections space to evolve, ignite new questions, or even serve as a picture of a past that feels a lot like the past, despite being so recent. Similar to the Bad/Good Aura, since late 2025 we emphasize human-only writing and GenAI tools only for brainstorming (if at all).
Our Shelves are:
Ignition, Shelf #3, April, 2026
(Ignition Selves #1 and #2 afterwards):
1. Technical, cold terms
Have you ever wondered whether the most harmful effects of GenAI tools are wrapped in technical, cold terms so that we don’t really understand them?
Why does it seem to be an imbalance between AI Leadership and “Innovation”, and human cognitive capital and resilience gained?
Will GenAI systems sustain aspects that function like a virus that neutralizes (or anesthetizes) its host, or may it change to at least keep the host cognitively alive just enough for a never-ending bite?
2. Policy makers.
I woke up with an itchy question, let’s see if it itches you as well. What happens if regulators or policymakers use GenAI tools to summarize or understand the content they are supposed to regulate? Doesn’t that mean that regulators are accessing (and assessing) the issues filtered by the lenses of tech interests?
Is that the new definition of “neutrality”?
Just the thought of that makes me bloody itchy. What about you? I know users can customize their tools, but I mean, those tools’ boundaries are still defined by tech companies, which have their own interests at play.
So, here is a question to ask your regulators: “Are you relying on GenAI tools to regulate us? If so, why?”
3. The only “alive” human traces.
After reading an article that “Lab-Grown Brains Growing More Powerful”, my little brain could not get away from three questions (if I am ignoring most of my ethical concerns about that kind of research):
a) What if humans as autonomous and individual creatures become obsolete, whereas humans as biological material and data become the only “alive” human traces in the world? (E.g., machine-human hybridization.)
b) At what point, if any, should lab-grown brain organoids be considered morally significant, especially if they develop complex neural activity?
c) Are humans like nonsense rebellious characters who will never stop playing with the world until it kills itself? Why don’t we act as Bob Sinclair’s rebel song (“World Hold On”) instead?
Why did you use “it” in the sentence above? It is there on purpose, as if humans were not really “humans” in the end, as if this nonsense search (AGI, AI race, and dangerous bioengineering) made us not “humans.” Since, we already have the Technology to end a lot of suffering and solve many societal problems, but instead of doing so, we keep this endless, restless search.
(Ignition Self #1 afterwards):
1. The secret sauce hypothesis.
What if a tech company poses itself as having developed an artificial general intelligence (AGI) system but actually hides a secret “human sauce”? A human ventriloquist, a conductor, in symbiosis with a machine in a fashion somewhat similar to this and those cases?
Doesn’t that trigger bells similar to The Mechanical Turk or the Automaton Chess Player? (May be of interest: the “Maelzel’s Chess Player” essay.)
What would it mean if the first “AGI” was part-human (is that AGI)?
What would that mean for the AGI race? Is a human-in-the-loop key to achieving AGI, or would you feel more comfortable if it were? If the “winner” of the AGI race offers a human-AI combo… did they really win?
2. AI-hygiene module.
Once GenAI becomes a daily cognitive companion at work, close to a human-machine hybrid, how long until companies start embedding “mental antivirus checks” into their infrastructure?
Things like: “look for signs such as getting stuck in loops, feeling overly confident because of an AI suggestion, or mistaking an AI’s tone for real emotions or agency”… the corporate equivalent of “please complete your annual cybersecurity training.”
And if you trigger one? A notification telling you that you must retake your “AI-hygiene module.”
If (when) this happens, I hope such training will foster the human ability to laugh at themselves and celebrate that “feeling silly” is a beautiful part of being human.
Then, corporate gossip may include: how many AI-hygiene modules were you required to take this month? As well as comments around an “AI-hygiene-free days counter”, similar to an accident-free days counter.
3. Human–AI Cognitive Signature Spill.
A reflection emerging in this space is what Eliott calls Human–AI cognitive Signature Spill: the idea that a person’s way of thinking (e.g., tone, rhythm, ambiguity thresholds, affective residues) may begin to imprint onto GenAI systems through repeated interaction, often without awareness or consent. Not in the sense of giving the AI your personal data, but in the sense of your mindset gradually shaping how the AI responds to you over time.
Once that becomes “commonplace”… what does privacy mean when what is being caught is not necessarily what we say, but how we think?
This adds an interesting dimension to data protection, neuro-rights, cognitive liberty, and transparency discussions. Thus, how to revisit, or even redefine these terms, when cognitive signature modeling and replication turn into an emerging reality?
4. Hold the nothing.
Why does it feel as if we are slowly killing our inner world? The Neverending Story carries 80s nostalgia, but also a scream of what we seem to be losing:
imagination, the ability to metabolize pain (the horse’s death), and the strength to face “the Nothing, the emptiness that is left”.
Sebastian’s brief conversation with his father is especially striking: a moment where a child is treated as someone capable of agency and emotional autonomy.
In a moment when GenAI is rapidly merging into daily routines and the Humanities are being undermined with alarming ease, we should ponder: how do we stop societal algorithmization from collapsing the human inner world? (A question the Empress would likely ask.)
If we lose the ability to really feel a piece of music, or to understand its deeper social and political meanings, what part of us is left? In other words, if we atrophy our ability to listen to a rap song and feel it in the stomach while also understanding its profound social, political, and structural critique, what exactly will remain of us?
A weakened inner world is easier to control and, as the movie’s wolf warns, very convenient for forces that thrive on emptiness. How is your ability to hold the nothing?
1. GenAI Tools and Visibility.
When reflecting on the current state of GenAI, and how close (or far) we are from AGI (Artificial General Intelligence), are we evaluating what is public or what is protected?
Are we assessing based on publicly available models (e.g., GPT, Claude, Gemini) or are we accounting for internal research-grade systems with restricted access?
Do we measure progress based only on the tools available to the public, or do we consider the possibility that more powerful systems are already in play, although not visible to us?
This gap raises important questions about benchmarking, transparency, and the limits of open-source discourse.
How should public knowledge evolve when the frontier may already have moved beyond view? What kind of epistemic illusion is reinforced if we keep assuming the “state of the art” is simply what we can access?
2. Lever Screens.
Do GenAI tools resemble a Skinner Box or remind you of the Addiction by Design book?
3. Rupture-Averse Scaffold and GenAI tools.
This is not to diminish the relevance but to reflect on the architecture: what if GenAI systems are structured to prioritize stability, user comfort, reflexive closure, and ambiguity aversion?
What if GenAI scaffolding limits epistemic bridges by minimizing uncertainty, paradox, and recursive collapse? What if it enforces engagement loops and optimizes for coherence and predictability, preventing systems from venturing into anti-structure cognition or discomfort-driven intelligence?
And what if those very approaches contribute to the uniformization of thought templates and language, ultimately creating more predictable users and thus making them more exploitable within digital infrastructures? (What if, in the long run, the most morally alive agents may be the ones whose bodies do not calm down when the group smiles?)
4. Pretended Comprehension.
If a machine’s output starts to get beyond the scope of human understanding, but we are still transitioning between being able to understand and not, transitioning to AI surpassing human comprehension (intelligence explosion), will humans pretend to understand?
Will we reach a tech version of The Emperor’s New Clothes, in which we pretend to understand something we are so ambiguous about? In the push to keep up, might we stop asking what is being said and start performing comprehension to fit?
5. What do we call intelligence?
When thinking about GenAI tools, a response to anthropomorphic impulses can focus not only on what they do and their effect on humans, but most importantly, on what we are willing to call intelligence.
What if we begin to mistake surface fluency for depth, and derivational mimicry for conceptual force? As language models grow more adept at reproducing human-like forms, we risk confusing mirroring with presence. But mirrors do not feel the contour of the face they reflect: they play with light, not meaning.
6. Uncomfortable Questions.
Today, a few years after Amy Webb’s 16 Uncomfortable Questions Everyone Needs to Ask About Artificial Intelligence, how do you feel about the AI landscape? Did you have a different opinion back in 2019?
- How to cite us: AI Aura team, “Ignition”, Shelf number, AI Aura
- We are happy to chat! Email us at: aiaura@grinnell.edu
- We use GenAI tools here, but not blindly (more info on our homepage).